Speakers have been shown to find sentences with erroneous agreement acceptable under certain cond... more Speakers have been shown to find sentences with erroneous agreement acceptable under certain conditions. This so-called agreement attraction effect has also been found in genitive-possessive structures such as "the teacher's brother" in Turkish (Lago et al., 2019), which is in contrast to its absence in similar constructions in English (Nicol et al., 2016). It has been hypothesized that this discrepancy is a result of the association between genitive case marking and subjecthood in Turkish. We test an alternative explanation according to which Lago et al.'s findings are due to a potential confound in their experiment, as all subject head nouns were locally ambiguous between possessive and accusative case. The results of our speeded acceptability judgment experiment suggest that the presence of case ambiguity does not affect agreement attraction, and thus suggests that genitive NPs may function as attractors in Turkish due to the association between genitive case and subjecthood.
This paper deals with the suffix-CIk in Turkish, and argues that the suffix-CIk has an augmentati... more This paper deals with the suffix-CIk in Turkish, and argues that the suffix-CIk has an augmentative meaning, as well as a diminutive meaning which has been mentioned in the literature extensively. The different meanings of the suffix-CIk are due to two different structures that I propose. While the structure of the diminutive is made of one single morpheme-CIk, that of the augmentative suffix is composed of two morphemes:-CI and-k. I offer phonotactic and morphological evidence that shows-k has its own life. I also show strict categorical selection properties for the diminutive and augmentative suffix: The augmentative one can only be attached to adjectival bases that express negative orderings. The diminutive-CIk, on the other hand, attaches to nouns. I capture this contrast using the nanosyntax model of spellout. Keywords. adjectives; augmentative; scalar; gradability; nanosyntax 1. Introduction. I start with some background on the suffix-CIk in Turkish. 1 Consider the following data: BASE-CIk FORM CONTRIBUTION STRESS MOVE K-DELETION * I would like to thank Aslı Göksel for helping me enormously with the topic itself and Carmen Savu for bearing with my questions about nanosyntax. I also want to thank Antonio Fabregas and Pavel Caha for their discussion. Lastly, I thank the anonymous reviewers on Tu+5, WoSSP 16, and AIMM4.
This paper takes two challenging characteristics of the Turkish case system and shows that a nano... more This paper takes two challenging characteristics of the Turkish case system and shows that a nanosyntactic analysis can cover both. The first puzzle is that some cases, namely ACC and GEN, in Turkish show alternations between specific and non-specific forms, while other cases like DAT and INS do not. The second puzzle concerns containment relations in morphology. Caha (2009) proposes that cases stand in a containment relation. In some languages like Estonian, Tocharian, and Vlax Romani, the ACC form serves as the foundation of the oblique cases. The puzzle is that in Turkish, the morphological containment holds only for ACC and GEN, but not for ACC and the other obliques. The comparison leads us to expect that the INS in Turkish could be *adam-ı-la, with the ACC marker to the left of-la. Interestingly, this expectation fails precisely in those cases which do not distinguish specific and non-specific forms. We propose a solution to both of these puzzles within the Nanosyntactic framework. The main idea is that Turkish nouns and cases can be composed of smaller, sub-morphemic features. These features allow specificity information to be encapsulated within the noun itself, rather than the case as previously suggested byÖztürk (2005).
Introduction. This paper presents a preliminary semantic analysis for the internally complex form... more Introduction. This paper presents a preliminary semantic analysis for the internally complex formative-MIŞ CASINA in Turkish which functions as a hypothetical comparison marker, as shown in (1). 1,2 We set aside its internal complexity and gloss it as HCM. 3 (1) Can Can arka-sın-dan [behind-POSS.3SG biri somebody koval-ıyor-muşcasına chase-IMPF]-HCM hızlı fast yürü-dü. walk-PST Lit: 'Can walked fast as if [somebody was chasing him].' Previous research on hypothetical comparative markers (also referred as comparative complementizers) is very limited with respect to the cross-linguistic typology. To the best of our knowledge, existing research on this topic only covers American English (Bledin & Srinivas 2019), Canadian English (Brook 2014), and German (Bücking 2017). Consequently, the analyses provided in the literature only address the empirical observations made for hypothetical comparison constructions in these languages. All available semantic analyses of hypothetical comparison make use of event semantics and invoke the notion of event similarity (Umbach & Gust 2014). Consider the following example from Bledin & Srinivas (2019) and the paraphrase they provide for it, which should suffice to illustrate the type of analysis they argue for. (2) a. Pedro danced as if he was possessed by the demons part that contributes hypothetical comparison * We thank Furkan Atmaca, Elena Guerzoni, and DenizÖzyıldız for their comments and contributions in the early forms of this work. We would also like to thank the anonymous reviewers on Tu+6.
This paper presents the first treebank for the Laz language, which is also the first Universal De... more This paper presents the first treebank for the Laz language, which is also the first Universal Dependencies Treebank for a South Caucasian language. This treebank aims to create a syntactically and morphologically annotated resource for further research. We also aim to document an endangered language in a systematic fashion within an inherently cross-linguistic framework: the Universal Dependencies Project (UD). As of now, our treebank consists of 576 sentences and 2,306 tokens annotated in light with the UD guidelines. We evaluated the treebank on the dependency parsing task using a pretrained multilingual parsing model, and the results are comparable with other low-resourced treebanks with no training set. We aim to expand our treebank in the near future to include 1,500 sentences. The bigger goal for our project is to create a set of treebanks for minority languages in Anatolia.
This paper takes two challenging characteristics of the Turkish case system and shows that a nano... more This paper takes two challenging characteristics of the Turkish case system and shows that a nanosyntactic analysis can cover both. The first puzzle is that some cases, namely ACC and GEN, in Turkish show alternations between specific and non-specific forms, while other cases like DAT and INS do not. The second puzzle concerns containment relations in morphology. Caha (2009) proposes that cases stand in a containment relation. In some languages like Estonian, Tocharian, and Vlax Romani, the ACC form serves as the foundation of the oblique cases. The puzzle is that in Turkish, the morphological containment holds only for ACC and GEN, but not for ACC and the other obliques. The comparison leads us to expect that the INS in Turkish could be *adam-ı-la, with the ACC marker to the left of -la. Interestingly, this expectation fails precisely in those cases which do not distinguish specific and non-specific forms. We propose a solution to both of these puzzles within the Nanosyntactic fram...
In this paper, we introduce the resources that we developed for Turkish dependency parsing, which... more In this paper, we introduce the resources that we developed for Turkish dependency parsing, which include a novel manually annotated treebank (BOUN Treebank), along with the guidelines we adopted, and a new annotation tool (BoAT). The manual annotation process we employed was shaped and implemented by a team of four linguists and five Natural Language Processing (NLP) specialists. Decisions regarding the annotation of the BOUN Treebank were made in line with the Universal Dependencies (UD) framework as well as our recent efforts for unifying the Turkish UD treebanks through manual re-annotation. To the best of our knowledge, BOUN Treebank is the largest Turkish treebank. It contains a total of 9,761 sentences from various topics including biographical texts, national newspapers, instructional texts, popular culture articles, and essays. In addition, we report the parsing results of a state-of-the-art dependency parser obtained over the BOUN Treebank as well as two other treebanks in Turkish. Our results demonstrate that the unification of the Turkish annotation scheme and the introduction of a more comprehensive treebank lead to improved performance with regards to dependency parsing.
Uploads
Papers by Utku Türk